DNA Sequences Useful for the Synthesis of Carotenoids

The present invention relates to DNA sequences which are useful for the synthesis of carotenoids such as lycopene, .beta.-carotene, zeaxanthin or zeaxanthin-diglucoside.

The present invention also relates to processes for producing such carotenoid compounds.

Carotenoids are widely distributed over green plants. They are yellow-orange-red lipids which are also present in some mold, yeast and so forth, and have recently received increased attention as natural coloring materials for foods. Among these carotenoids, .beta.-carotene is a typical one, which is used as a coloring materials and as a precursor of vitamin A in mammals as well. It is also expected to be useful for preventing cancer [see, for example, SHOKUHIN TO KAIHATSU (Foods and Development), 24, 61-65 (1989)]. Since carotenoids such as .beta.-carotene are widely distributed over green plants, plant tissue culture has been examined as a means to produce carotenoids in a large amount without the influence of natural environment [see, for example, Plant Cell Physiol., 12, 525-531 (1971)]. Examination has also been made to search for a microorganism such as mold, yeast or green algae with a high carotenoid producing activity and to produce carotenoids in a large amount using such microorganism (see, for example, The Abstract of Reports in the Annual Meeting of NIPPON HAKKO KOGAKU-KAI of 1988, page 139). However, neither of these is successful at present in producing .beta.-carotene at a good productivity which exceeds synthetic methods in commercial production. It would be very useful to obtain a gene group which participates in the biosynthesis of carotenoids, because it will be possible to produce carotenoids in a larger amount by introducing and expressing appropriate gene(s) among such group in an appropriate host such as a plant tissue cultured cell, a mold, a yeast or the like which originally produces carotenoids. This will lead to a .beta.-carotene production superior to synthetic methods, or to the production of useful carotenoids other than .beta.-carotene in a large amount.

Furthermore, the gene group participating in the biosynthesis of carotenoids would make it possible to synthesize carotenoids in a cell or an organ which originally has no carotenoid producing ability, which will add new values to the organism. For example, several reports have recently been made with reference to creating flower colors not found in nature by the genetic manipulation of flowering plants [see, for example, Nature, 330, 677-678 (1987)]. Flower colors are given by pigments such as anthocyanin and carotenoids. Anthocyanin is responsible for flower colors of red-violet-blue, and carotenoids are responsible for flower colors of yellow-orange-red. Genes encoding enzymes involved in anthocyanin synthesis have been elucidated, which were used for creating a new flower color in the aforementioned reports. On the other hand, there are many flowering plants having no bright yellow flower due to lack of carotenoid synthesis in petals (e.g. petunia, saintpaulia (African violet), cyclamen, Primula malacoides, etc.). Expression of appropriate reconstructed gene(s) for carotenoid biosynthesis in petals would give yellow flowers to these flowering plants.

However, enzymes for synthesizing carotenoids or genes coding for them are scarcely elucidated at present. The nucleotide sequence of the gene group participating in the biosynthesis of a kind of carotenoids has been elucidated lately in a photosynthetic bacterium Rhodobacter capsulatus [Mol. Gen. Genet., 216, 254-268 (1989)]. This bacterium, however, synthesizes an acyclic xanthophyll spheroidene via neurosporene without cyclization and thus cannot synthesize general carotenoids such as lycopene, .beta.-carotene and zeaxanthin.

There are prior arts with reference to yellow pigments or carotenoids of Erwinia species disclosed in J. Bacteriol., 168, 607-612 (1986), J. Bacteriol., 170, 4675-4680 (1988) and J. Gen. Microbiol., 130, 1623-1631 (1984). The first one of these references discloses the cloning of a gene cluster coding for yellow pigment synthesis from Erwinia herbicola Eho 10 ATCC 39368 as a 12.4 kilobase pairs (kb) fragment. There is no information on the nucleotide sequence of the 12.4 kb fragment. The second literature discloses a yellow pigment synthesized by the cloned gene cluster, which is indicated to be a carotenoid by the analysis of its UV-visible spectrum. The last literature indicates that the genes participating in the production of yellow pigments in Erwinia uredovora 20D3 ATCC 19321 are present on a 260 kb large plasmid from the observation that the yellow pigments are not produced after curing the large plasmid, and further discloses that the pigments are carotenoids flora the analysis of their UV-visible spectrum.

However, the chemical structures of the carotenoids and their precursors produced by the Erwinia species, enzymes participating in their synthesis or the nucleotide sequences of the genes encoding these enzymes have been unknown. Very recently, an international patent application was published (WO 91/13078, published on Sep. 5, 1991) which discloses the isolation and nucleotide sequences of the carotenoid synthesis genes of Erwinia herbicola Eho 10 and their expression in several organisms.

The object of the present invention is to provide DNA sequences which are useful for the synthesis of carotenoids such as lycopene, .beta.-carotene, zeaxanthin or zeaxanthin-diglucoside; DNA sequences encoding casotenoid biosynthesis enzymes.

In other words, the DNA sequences useful for the synthesis of carotenoids according to the present invention are the DNA sequences 1-6 described in the following (1)-(6).

(1) DNA sequence 1: a DNA sequence encoding a polypeptide which has an enzymatic activity for converting farnesyl pyrophosphate into geranylgeranyl pyrophosphate and whose amino acid sequence corresponds substantially to the amino acid sequence of SEQ ID NO:1;

(2) DNA sequence 2: a DNA sequence encoding a polypeptide which has an enzymatic activity for converting zeaxanthin into zeaxanthin-diglucoside and whose. amino acid sequence corresponds substantially to the amino acid sequence of SEQ ID NO:2;

(3)DNA sequence 3: a DNA sequence encoding a polypeptide which has an enzymatic activity for converting lycopene into .beta.-carotene and whose amino acid sequence corresponds substantially to the amino acid sequence of SEQ ID NO:3;

(4)DNA sequence 4: a DNA sequence encoding a polypeptide which has an enzymatic activity for converting phytoene into lycopene and whose amino acid sequence corresponds substantially to the amino acid sequence of SEQ ID NO:4;

(5)DNA sequence 5: a DNA sequence encoding a polypeptide which has an enzymatic activity for converting geranylgeranyl pyrophosphate into phytoene and whose amino acid sequence corresponds substantially to the amino acid sequence of SEQ ID NO:5; and

(6)DNA sequence 6: a DNA sequence encoding a polypeptide which has an enzymatic activity for converting .beta.-carotene into zeaxanthin and whose amino acid sequence corresponds substantially to the amino acid sequence of SEQ ID NO:6.

Another object of the present invention is to provide processes for producing carotenoid compounds.

More specifically, the present invention also provides a process for producing a carotenoid or a precursor compound which is selected from the group consisting of geranylgeranyl pyrophosphate, phytoene, lycopene, .beta.-carotene, zeaxanthin and zeaxanthin-diglucoside, which comprises transforming a host with at least one of the DNA sequences 1-6 described above and culturing the transformant.

The successful acquirement of the gene group useful for the synthesis of carotenoids such as lycopene, .beta.-carotene, zeaxanthin, zeaxanthin-diglucoside or the like (gene group encoding the biosynthetic enzymes of carotenoids) according to the present invention has made it possible to produce useful carotenoids in large amounts, for example, by employing an appropriate plant tissue cultured cell, a microorganism or the like transformed with a plasmid capable of expressing the gene(s) to a high level. The present invention has also made it possible to synthesize carotenoids in cells or organs which produce no carotenoid by creating a plasmid in which the gene(s) can be expressed in the target cell or organ and transforming a suitable host with this plasmid.

The DNA sequences according to the present invention are the aforementioned DNA sequences 1-6, that is, genes encoding the polypeptides of respective enzymes which participate in the biosynthesis reaction of carotenoids, particular, for example, such polypeptides in Erwinia uredovora 20D3 ATCC 19321.

A variety of gene groups containing a combination of the DNA sequences 1-6 can be expressed in a microorganism, a plant or the like to afford them the biosynthesis ability of carotenoids such as lycopene, .beta.-carotene, zeaxanthin, zeaxanthin-diglucoside or the like. The DNA sequences may be present on a single DNA strand (genomic DNA or a plasmid), or on multiple separate DNA strands genomic DNA and/or plasmids) individually or in groups.

The aforementioned gene groups encode a plurality of enzymes participating in the production of carotenoids. It can be incorporated into a proper vector and then introduced into a suitable host to create a transformant. Culturing the transformant will result in the production of the plurality of enzymes participating in carotenoid formation, which will synthesize the carotenoids in the transformant.

The DNA sequence of SEQ ID NO:13, which is an example according to the present invention, is acquired from Erwinia uredovora 20D3 ATCC 19321 and shows, as illustrated in the experimental example below, no homology by DNA-DNA hybridization with the DNA strand containing the gene group for synthesizing the yellow pigments of Erwinia herbicola Eho 10 ATCC 39368 (see Background of the Invention described above).

The DNA sequences of the present invention are the DNA sequences 1-6 which contain nucleotide sequences encoding a polypeptide whose amino acid sequence corresponds substantially to one of the amino acid sequences of SEQ ID NO:1 to 6. In the present invention, the "DNA sequence" is defined by the amino acid sequence of the polypeptide encoded by the DNA sequence so that each "DNA sequence" has a definite length according to the length of the polypeptide. Although the "DNA sequence" contains a gene encoding an enzyme and is useful for biotechnological production of the enzyme, such biotechnological production cannot be performed by the "DNA sequence" of the definite length alone but in a state where other DNA sequences with a proper length are linked to the 5'-upstream and/or the 3'-downstream of the "DNA sequence". The term "DNA sequence" in the present invention therefore includes, in addition to those having the definite length, a linear DNA strand or a circular DNA strand containing this "DNA sequence" as its part.

One of the typical forms of the DNA sequences of the present invention is a plasmid which comprises the DNA sequences as its part, which may be present in a host such as E. coli. A preferable plasmid according to the present invention is a replicable plasmid that is maintained stably in the host carrying the DNA sequences of the present invention, as a passenger or a foreign gene, linked to a promoter(s) (containing the ribosome-binding site in the case of a prokaryotic host). As the plasmid vector and the promoter, an appropriate combination of those which are well-known can be used.

The polypeptides encoded by the DNA sequences of the present invention have an amino acid sequence which corresponds substantially to one of the amino acid sequences of SEQ ID NO:1 to 6. Here, in the six polypeptides of SEQ ID NO:1 to 6 (i.e. six enzymes participating in the formation of carotenoids), some of the amino acids can be deleted or substituted or some amino acids can be added or inserted, so long as each polypeptide has the aforementioned enzymatic activity in terms of substrate-product conversion. This is represented by the expression "whose amino acid sequence corresponds substantially to ..." in the claims. For example, polypeptides with a deletion of the first amino acid (Met) from the polypeptides of SEQ ID NO:1 to 6 are included in such deleted polypeptides.

The amino acid sequences of SEQ ID NO:1 to 6 have not been known before.

The DNA sequences encoding these enzymes are those having one of the nucleotide sequences of SEQ ID NO: 7 to 12 and degenerative isomers thereof, or those having the nucleotide sequences corresponding to an aforementioned alteration of the amino acid sequence of these enzymes and degenerative isomers thereof. The term "degenerative isomer" means a DNA sequence which is different only in degenerative codons and codes for the same polypeptide. The 5'-upstream and/or the 3'-downstream of the DNA sequences of the present invention may further have a DNA sequence with a certain length as non-coding regions.

The gene groups used for the synthesis of carotenoid, whose typical examples are described in the following (1)-(4), comprise a plurality of the aforementioned DNA sequences 1-6. Each gene group encodes a plurality of enzymes which participate in the production of carotenoids from their substrates.

One example of a gene group used for the synthesis of lycopene, a red carotenoid, comprises the aforementioned DNA sequences 1, 4 and 5. The DNA sequences may be present on one DNA strand, or on separate multiple DNA strands individually or in groups.

In the case that a plurality of DNA sequences are present on one DNA strand, the arrangement order and direction of the DNA sequences may be optional provided that the genetic information is expressed, i.e. the genes are in a state capable of being transcribed and translated appropriately in the host.

The biosynthetic pathway of lycopene in E. coli is explained as follows: geranylgeranyl pyrophosphate which is a substrate originally present in E. coli is converted into phytoene by the enzyme encoded by the DNA sequence 5 and the phytoene is further convened into lycopene by the enzyme encoded by the DNA sequence 4. The enzyme encoded by the DNA sequence I converts farnesyl pyrophosphate into geranylgeranyl pyrophosphate and thus is necessary to increase the production of lycopene (see FIG. 2).

Lycopene is a red carotenoid. It is present in a large amount in the fruits of watermelon or tomato and is highly safe to eat. In this connection, the lycopene synthesized by the DNA sequences of the present invention in the experimental example described below had the same stereochemistry as the lycopene present in these plants.

The gene group is typically on a plasmid with a stop codon for each gene, which plasmid may or may not be present in a host such as E. coil. A preferable plasmid according to the present invention is a replicable plasmid maintained stably in the host carrying the gene group, as a passenger or a foreign gene, linked to a promoter(s) (containing ribosome-binding sites in the case of a procaryote). In procaryotes such as E. coil or Zytnomonas species, a single promoter can be used to express all the DNA sequences, or alternatively each DNA sequence may have its own promoter. In the case of eucaryotes such as a yeast or a plant, each DNA sequence preferably has its own promoter.

One example of a gene group used for the synthesis of .beta.-carotene, a yellow-orange carotenoid, comprises the aforementioned DNA sequences 1, 3, 4 and 5. In other words, the gene group used for the synthesis of .beta.-carotene is formed by adding the DNA sequence 3 to the gene group used for the synthesis of lycopene comprising the DNA sequences 1, 4, and 5. The DNA sequences may be present on one DNA strand, or on separate multiple DNA strands individually or in groups.

In the case that a plurality of DNA sequences are present on one DNA strand, the arrangement order and direction of the aforementioned DNA sequences may be optional provided that the genetic information is expressed, i.e. the genes are in a state capable of being transcribed and translated appropriately in the host.

The biosynthetic pathway of .beta.-carotene in E. coli is explained as follows: geranylgeranyl pyrophosphate which is a substrate originally present in E. coli is converted into phytoene by the enzyme encoded by the DNA sequence 5 the phytoene is further converted into lycopene by the enzyme encoded by the DNA sequence 4, and the lycopene is further convened into .beta.-carotene by the enzyme encoded by the DNA sequence 3. The enzyme encoded by the DNA sequence 1 converts farnesyl pyrophosphate into geranylgeranyl pyrophosphate and thus is necessary to increase the production of lycopene (see FIG. 2).

.beta.-carotene is a typical carotene with a color spectrum ranging from yellow to orange, and it is the orange pigment present in a large amount in the roots of carrot or green leaves of plants and is highly safe to eat. The utility of .beta.-carotene has already been described in Background of the Invention. In this connection, the .beta.-carotene synthesized by the DNA sequence of the present invention in the experimental example described below had the same stereochemistry as the .beta.-carotene present in the roots of carrot or green leaves of plants.

The gene group is typically on a plasmid as in the same manner as defined (1):

One example of a gene group used for the synthesis of zeaxanthin, another yellow-orange carotenoid, comprises the aforementioned DNA sequences 1, 3, 4, 5 and 6. In other words, the DNA sequence used for the synthesis of zeaxanthin is formed by adding the DNA sequence 6 to the gene group used for the synthesis of .beta.-carotene comprising the DNA sequences 1, 3, 4 and 5. The DNA sequences may be present on one DNA strand, or on separate multiple DNA strands individually or in groups.

In the case that a plurality of DNA sequences are present on one DNA strand, the arrangement order and direction of the aforementioned DNA sequences may be optional provided that the genetic information is expressed, i.e. the genes are in a state capable of being transcribed and translated appropriately in the host.

The biosynthetic pathway of zeaxanthin in E. coli is explained as follows: geranylgeranyl pyrophosphate which is a substrate originally present in E. coli is convened into phytoene by the enzyme encoded by the DNA sequence 5 the phytoene is then convened into lycopene by the enzyme encoded by the DNA sequence 4, and the lycopene is further convened into .beta.-carotene by the enzyme encoded by the DNA sequence 3, and finally the .beta.-carotene is convened into zeaxanthin by the enzyme encoded by the DNA sequence 6. The enzyme encoded by the DNA sequence 1 converts farnesyl pyrophosphate into geranylgeranyl pyrophosphate and thus is necessary to increase the production of lycopene (see FIG. 2).

Zeaxanthin is a xanthophyll with a color spectram ranging from yellow to orange, and it is the yellow pigment which is present in the seed of maize and is highly safe for food. Zeaxanthin is contained in feeds for hen or colored carp and is an important pigment source for coloring them. In this connection, the zeaxanthin synthesized by the DNA sequences of the present invention in the experimental example described below had the same stereochemistry as the zeaxanthin described above.

The gene group is typically on a plasmid as in the same manner as defined in (1):

One example of a gene group used for the synthesis of zeaxanthin-diglucoside, yet another yellow-orange carotenoid, comprises the aforementioned DNA sequences 1-6. In other words, the gene group used for the synthesis of zeaxanthin-diglucoside is formed by adding the DNA sequence 2 to the gene group used for the synthesis of zeaxanthin comprising the DNA sequences 1, 3, 4, 5 and 6. The DNA sequences may be present on one DNA strand, or on separate multiple DNA strands individually or in groups.

In the case that a plurality of DNA sequences are present on one DNA strand, the arrangement order and direction of the aforementioned DNA sequences may be optional provided that the genetic information is expressed, i.e. the genes are in a state capable of being transcribed and translated appropriately in the host.

The gene group is typically on a plasmid as in the same manner as defined in (1).

The biosynthetic pathway of zeaxanthin-diglucoside in E. coli is explained as follows: geranylgeranyl pyrophosphate which is a substrate originally present in E. coli is converted into phytoene by the enzyme encoded by the DNA sequence 5 the phytoene is then converted into lycopene by the enzyme encoded by the DNA sequence 4, and the lycopene is further converted into .beta.-carotene by the enzyme encoded by the DNA sequence 3, the .beta.-carotene is then converted into zeaxanthin by the enzyme encoded by the DNA sequence 6, and the zeaxanthin is finally converted into zeaxanthin-diglucoside by the enzyme encoded by the DNA sequence 2. The enzyme encoded by the DNA sequence I converts farnesyl pyrophosphate into geranylgeranyl pyrophosphate and thus is necessary to increase the production of lycopene (see FIG. 2).

Zeaxanthin-diglucoside is a carotenoid glycoside with a high water solubility which is sufficiently soluble in water at room temperature to exhibit a clear yellow color. Carotenoid pigments are generally hydrophobic and thus limited in their use as natural coloring materials in foods or the like. Zeaxanthin-diglucoside overcomes this defect. It has been isolated from edible plant saffron (Croccus sativus; Pure & Appl. Chem., 47, 121-128 (1976)), thus its safety for food apparently being continued. Zeaxanthin-diglucoside is therefore desirable as a yellow natural coloring material of foods or the like. There has been heretofore no reports with reference to the isolation of zeaxanthin-diglucoside from microorganisms.

When carotenoid pigments such as lycopene, .beta.-carotene, zeaxanthin and zeaxanthin-diglucoside are to be produced in E. coli, the aforementioned DNA sequences 1, 4 and 5, the DNA sequences 1, 3, 4 and 5, the DNA sequences 1, 3, 4, 5 and 6, and the DNA sequences 1-6 are required, respectively. However, when a host other than E. coli, particularly one which is capable of producing carotenoids is used, it is highly possible that the host contains an intermediate compound further downstream in the carotenoid biosynthesis, so that not all of the aforementioned DNA sequences 1, 4 and 5 (for the production of lycopene), the DNA sequences 1, 3, 4 and 5 (for the production of 13-carotene), the DNA sequences 1, 3, 4, 5 and 6 (for the production of zeaxanthin), or the DNA sequences 1-6 (for the production of zeaxanthin-diglucoside) may be required.

That is to say, only the DNA sequence(s) participating in the formation of the aimed carotenoid pigment from the furthest downstream intermediate compound present in the host may be necessary in these cases. Thus, when lycopene is intended to be produced in a host in which phytoene is already present, it will be sufficient to use only the DNA sequence 4 among the DNA sequences 1, 4 and 5.

It is also possible to make a host produce phytoene, a carotenoid pigment related compound, by using the DNA sequence 5 alone or, preferably, the DNA sequences 1 and 5 of the present invention.

One method for acquiring the DNA sequences 1-6 which contain the nucleotide sequences coding for the aforementioned enzymes is the chemical polynucleotide synthesis of at least a part of their strands. However, taking into consideration that a number of nucleotides to be bonded is big, it would be more preferable than the chemical synthesis to acquire the DNA sequences from a DNA library of Erwinia uredovora 20D3 ATCC 19321 according to conventional methods in the field of genetic engineering, for example, hybridization with suitable probes.

Individual DNA sequences or a DNA sequence comprising all the DNA sequences can be obtained in these ways.

The aforementioned gone groups comprising a combination of the DNA sequences 1-6 can be constructed using the DNA sequences obtained as described above. The DNA sequences thus obtained contain genetic information for enzymes participating in the formation of carotenoids, so that by introducing them into an appropriate host by biotechnological techniques, it will be able to obtain a transformant producing the enzymes and thus carotenoid pigments or carotenoid pigment related compounds.

Plants and a variety of microorganisms, as far as a suitable host-vector system is available can be transformed by a vector comprising the aforementioned DNA sequences. However, the host is required to contain farnesyl pyrophosphate which is a substrate compound of the first enzyme of the carotenoid synthesis by the DNA sequences of the present invention, or an intermediate compound further downstream from it.

It is known that farnesyl pyrophosphate is synthesized by phenyltransferases such as farnesyl pryophosphate synthetase which are common enzymes in early steps of the biosynthesis of carotenoids as well as steroids and terpenes [J. Biochcm., 72, 1101-1108 (1972)]. Accordingly, even a cell which cannot synthesize carotenoids must contain farnesyl pyrophosphate if it synthesizes steroids or terpenes. Since every cell contains at least either steroids or terpenes, theoretically almost all hosts are capable of synthesizing carotenoids using the DNA sequences of the present invention as far as a suitable host-vector system is available. Host-vector systems are known, for example, for plants such as Nicotiana tabacura, Petunia hybrida and the like microorganisms such as bacteria, for example Escherichia coli, Zymornonas mobilis and the like, and yeasts, for example Saccharomyces cerevisiae and the like.

It is continued here for the first time that the genetic information on the DNA sequences of the present invention can be expressed in transformed microorganisms. However, the procedures or the methods for transformation (and the production of enzymes and thus carotenoid pigments or carotenoid pigment related compounds) per se are conventional in the fields of molecular biology, cell biology or genetic manipulation, and therefore the procedures not described in detail below can be performed according to these conventional techniques.

In order to express the genes on the DNA sequences of the present invention in a host, it is necessary to insert the genes into a vector to introduce it into the host. As the vector, any of various known vectors can be used such as pBI121 or the like for plants (Nicotiana tabacura, Petunia hybrida); pUC19, pACYC184 or the like for E. coli; pZA22 or the like for Zymomonas mobilis (see Japanese Patent Laid-Open Publication No. 228278/87); and YEp13 or the like for yeast.

Furthermore, it is necessary to transcribe the DNA sequences of the present invention into mRNA in order to express the genes on the DNA sequences in the host. For this purpose, a promoter, a signal for the transcription, is linked to the 5'-upstream of the DNA sequences of the present invention. A variety of promoters such as CaMV35S, NOS, TR1', TR2' (for plants); lac, Tc.sup.r, CAT, trp (for E. coli); Tc.sup.r, CAT (for Zymomonas mobilis); ADH1, GAL7, PGK, TRP1 (for yeast) and the like can be used in the present invention.

In the case of prokaryotic hosts, it is necessary to place a ribosome-binding site (SD sequence in E. coli) several base-upstream from the initiation codon (ATG).

The transformation of the host with the vector thus obtained can be conducted by any appropriate method which is conventionally used in the fields of genetic manipulation or cell biology. Appropriate publications or reviews can be referenced; for example, for the transformation of microorganisms, T. Maniatis, E. F. Fritsch and J. Sambrook: "Molecular Cloning A Laboratory Manual", Cold Spring Harbor Laboratory (1982).

The transformant is identical with the used host in its genotype, phenotype or bacteriological properties except for the new trait derived from the genetic information introduced by the DNA sequence of the present invention (that is, the production of enzymes participating in the carotenoid formation and the synthesis of carotenoids or the like by the enzymes), the trait derived from the used vector, and possible losses of traits corresponding to the deletion of a part of the genetic information which might be caused by a recombination with the vector. Escherichia coli JM109 (pCAR1) which is an example of the transformant according to the presefit invention is deposited as FERM BP-2377.

The transformants obtained as described above produce, mainly within the cell, enzymes participating in the carotenoid formation, and a variety of carotenoids or carotenoid pigment related compounds are synthesized by the enzymes.

Culture or the culturing condition of the transformant is essentially the same as those for the host used.

Carotenoids can be recovered by the methods, for example, illustrated in Experimental Examples 3 and 4 below.

Furthermore, enzyme proteins encoded by the DNA sequences of the present invention are produced mainly within the cell in the case of E. coli transformants and can be recovered by an appropriate method.

FIGS. 1(a) through 1(g) illustrate the KpnI-HindIII fragment which was acquired from Erwinia uredovora 20D3 ATCC 19321 and involved in the biosynthesis of carotenoids, that is the complete nucleotide sequence of the 6918 bp DNA sequence (SEQ ID NO:13) containing the DNA sequences SEQ ID NO:I to 6, and

FIG. 2 illustrates the function of the polypeptides encoded by the aforementioned DNA sequences 1-6.

FIG. 3 illustrates the structure of vector Y-crtBEIY.

All of the strains used in the following experiments are deposited in ATCC or other deposition organizations and are freely available.

Experimental Example 1: Cloning of a gene cluster participating in the biosynthesis of a yellow pigment (referred to hereinafter as yellow pigment-synthesizing gene cluster)

Total DNA was prepared from the cells of Erwinia uredovora 20D3 ATCC 19321 which had been proliferated until early-stationary phase in 100 ml of LB medium (1% tryptone, 0.5% yeast extract, 1% NaCl). Penicillin G (manufactured by Meiji Seika) was added to the culture medium to a concentration of 50 units/ml 1 hour before the harvest of the cells. The cells were harvested by centrifugation, washed with TES buffer (20 mM Tris, 10 mM EDTA, 0.1 M NaCl, pH 8), heat treated at 68.degree. C. for 15 minutes and suspended in Solution I (50 mM glucose, 25 mM Tfis, 10 mM EDTA, pH 8) containing 5 mg/ml of lysozyme (manufactured by Seikagaku Kogyo) and 100 .mu.g/ml of RNase A (manufactured by Sigma). The suspension was incubated at 37.degree. C. for 30 minutes to 1 hour, and pronase E (manufactured by K. aken Seiyaku) was added to a concentration of 250 .mu.g/ml before incubating at 37.degree. C. for further 10 minutes. Sodium N-lauroylsarcosine (manufactured by Nacalai tesque) was added to a final concentration of 1%, and the suspension was mixed well and incubated at 37.degree. C. for several hours. After several times of phenol/chloroform extractions, two volumes of ethanol was slowly added to recover the total DNA by winding around a glass stick. The DNA was rinsed with 70% ethanol and dissolved in 2 ml of TE buffer (10 mM Tris, 1 mM EDTA, pH 8) to give the total DNA preparation.

26 .mu.g of the total DNA was incubated with I unit of a restriction enzyme Sau3AI at 37.degree. C. for 30 minutes followed by a heat inactivation of the restriction enzyme at 68.degree. C. for 10 minutes. Sau3AI partial digestion fragments of about 40 kb were obtained under this condition. The DNA was recovered by ethanol precipitation and a half portion was mixed with 2.5 .mu.g of BamHI digested cosmid pJB8 treated with alkaline phosphatase and 0.2 .mu.g of the pJB8 SalI-BamHI right arm fragment (smaller fragment) recovered by a gel electrophoresis in a total amount of 40 .mu.l and was subjected to ligation reaction with T4 DNA ligase at 12.degree. C. for 2 days. The cosmid pJB8 had been previously purchased from Amersham. Restriction enzymes and enzymes used for genetic manipulation were purchased from Boehringer-Mannheim, Takara Shuzo or Wako Pure Chemical Industries. The ligated DNA was used for in vitro packaging with Gigapack Gold (manufactured by Stratagene) to give phage particles sufficient for constructing a cosmid library. Escherichia coil DH1 (ATCC 33849) cells were infected with the phage particles, diluted to several hundred colonies per plate, and were plated on LB plates, which were cultured at 37.degree. C. overnight and further at 30.degree. C. for 6 hours or more. E. coli transformants producing yellow pigments appeared in a proportion of one colony per about 1,100 colonies. The E. coli transformants producing yellow pigments contained plasmids in which 33-47 kb Sau3AI partial digestion fragments were inserted into pJB8.

One of the yellow pigment-synthesizing clones obtained above was arbitrarily chosen and its Sau3AI partial digestion fragment was subjected to another partial digestion with Sau3AI, ligated to the BamHI site of an E. coli vector pUC19 (purchased from Takara Shuzo), which was used to transform Escherichia coli JM109 (Takara Shuzo) in order to locate the yellow pigment-synthesizing gene cluster. Plasmid DNAs were prepared from 50 E. coli transformants producing yellow pigments which appeared on LB plates containing ampicillin, and analyzed by agarose gel electrophoresis. The smallest insert was 8.2 kb. The plasmid containing this 8.2 kb fragment was named pCAR1 and E. coli JM109 harboring this plasmid was named Escherichia coli JM109 (pCAR1). This strain produced the same yellow pigments as those of E. uredovora. The 8.2 kb fragment had a KpnI site near the end of the/ac promoter side and a HindIII site near the opposite end. pCAR1 was doubly digested with KpnI/HindIII (HindIII digestion was partial; the 8.2 kb fragment had two HindIII sites) and the KpnI-HindIII fragment (6.9 kb) was recovered by a gel electrophoresis and inserted between the KpnI and HindIII sites of pUC18 (this hybrid plasmid was named pCAR15). E. coli JM109 transformant carrying pCAR15 made a yellow colony and produced the same yellow pigments as those of E. uredovora. Accordingly, the genes required for the yellow pigment production were located on this KpnI-HindIII fragment (6.9 kb). In other words, the fragment carrying the yellow pigment-synthesizing genes was reduced to 6.9 kb in size.

Experimental Example 2: Analysis of the yellow pigment-synthesizing gene cluster

The complete nucleotide sequence of the 6.9 kb KpnI-HindIII fragment was determined by the kilo-sequence method using a deletion kit for kilo-sequence (manufactured by Takara Shuzo) and the dideoxy method according to Proc. Natl. Acad. Sci. USA, 74 5463-5467 (1977). The KpnI-HindIII fragment containing the yellow pigment-synthesizing genes was 6918 base pairs (bp) in length and its GC content was 54%. The complete nucleotide sequence is shown in FIGS. 1(a)-(g) (SEQ ID NO:13). The nucleotides are numbered as the KpnI site being number 1.

The HindIII side of the 6918 bp DNA fragment containing the yellow pigment-synthesizing genes (the 3' side of SEQ ID NO:13) was shortened using the deletion kit for kilo-sequence. A hybrid plasmid (designated pCAR25) carrying nucleotide 1-6503 of SEQ ID NO:13 was constructed by inserting the fragment obtained by deleting from the HindIII site to nucleotide position 6504 into pUC19. E. coli JM109 harboring pCAR25 [referred to hereinafter as E. coli (pCAR25)] made a yellow colony and produced the same yellow pigments as those of E. uredovora. The region from the nucleotide number 6504 to 6918 in SEQ ID NO:13 was therefore not required for the yellow pigment production. The region from the nucleotide number I to 6503 containing the yellow pigment-synthesizing genes was analyzed and six open reading flames (ORFs) were found in this region. These are: an ORF from the nucleotide number 225 to 1130 coding for a polypeptide with a molecular weight of 32,583 (ORF1, A - B in FIG. 1); an ORF from the nucleotide number 1143 to 2435 coding for a polypeptide with a molecular weight of 47,241 (ORF2, C - D in FIG. 1); an ORF from the nucleotide number 2422 to 3567 coding for a polypeptide with a molecular weight of 43,047 (ORF3, E - F in FIG. 1); an ORF from the nucleotide number 3582 to 5057 coding for a polypeptide with a molecular weight of 55,007 (ORF4, G - H in FIG. 1); an ORF from the nucleotide number 5096 to 5983 coding for a polypeptide with a molecular weight of 33,050 (ORF5, I - J in FIG. 1); and an ORF from the nucleotide number 6452 to 5928 coding for a polypeptide with a molecular weight of 19,816 (ORF6, K - L in FIG. 1. This ORF6 alone has the opposite orientation with the others). Each ORF had an SD (Shine-Dalgarno) sequence complementary to the 3'-region of 16S ribosomal RNA of E. coil at several nucleotides upstream of its initiation codon, which strongly suggested that polypeptidcs were in fact synthesized in E. coli by these six ORFs. This was confirmed by the in vitro transcriptio-translation experiments described below.

In vitro transcription-translation analyses were carried out with the prokaryotic DNA-directed translation kit of Amersham using either ScaI-digested plasmid pCAR25 (carrying ORF1 - ORF6) or ScaI-digested plasmids carrying one of the ORFs 1-6 (with the SD sequence) as the template DNA. The later plasmids were constructed by isolating DNA fragments carrying erich ORF using appropriate restriction enzymes and inserting them into pUC18 or pUC19 in a manner that the inserted fragment would be transcribed by a read-through transcription of the lac promoter. It was confirmed that the polypeptide bands corresponding to the aforementioned respective ORFs were detected among the transcription-translation products.

All of the six ORFs were necessary for the production of the same yellow pigments as those of E. uredovora as described below (Experimental Example 3, 4 and 5). ORF1, ORF2, ORF3, ORF4, ORF5 and ORF6 ,were designated as crtE, crtX, crtY, crtI, crtB and crtZ genes, respectively. These genes were previously referred to as zexA, zexB, zexC, zexD, zexE and zexF, respectively, in the priority Japanese applications and in the US application 07/519,011 to which the present application is accorded the benefit of an earlier filing date. However, since some of these genes, namely ORFs 1, 4 and 5, have a functional similarity to locus of Rhodobacter capsulatus, which had been already termed as crtE, crtI and crtB locus [Molec. Gen. Genet., 216, 254-268 (1989)], the gene names have been altered accordingly after the filing of these prior applications.

Total DNA of Erwinia herbicola Eho 10 ATCC 39368 was prepared in the same manner as in Experimental Example 1 (1). A 7.6 kb fragment containing the DNA sequence of SEQ ID NO:13 was cut out flora the hybrid plasmid pCAR1 by KpnI digestion and labeled with DNA labeling & detection kit (manufactured by Boehringer-Mannheim) by the DIG-ELISA method to give the probe DNA. Total DNAs (intact or KpnI digested) of E. herbicola Eho 10 ATCC 39368 and E. uredovora 20D3 ATCC 19321 were electrophoresed on an agarose gel and hybridized with this probe DNA using the aforementioned DNA labeling & detection kit. The probe DNA strongly hybridized with the total DNA of E. uredovora 20D3 ATCC 19321 but not at all with that of E. herbicola Eho 10 ATCC 39368. Also, the restriction map deduced from the DNA sequence of SEQ ID NO:13 was quite different from that reported in J. Bactefiol., 168, 607-612 (1986) for the E. herbicola DNA. It was concluded from these results that the DNA sequence of SEQ ID NO:13, that is, the DNA sequence useful for the synthesis of carotenoids according to the present invention, is not homologous with the DNA sequence containing the yellow pigment-synthesizing genes of E. herbicola Eho 10 ATCC 39368.

Experimental Example 3: Analysis of yellow pigments

E. coli (pCAR25) produced the same yellow pigments as those of E. uredovora 20D3 ATCC 19321 and E. herbicola Eho 10 ATCC 39368, and their yield was 5 times higher than that of the former and 6 times higher than that of the latter (per dry weight). Cells harvested from 8 liters of 2.times.YT medium (1.6% tryptone, 1% yeast extract, 0.5% NaCl) were extracted once with 1.2 liter of methanol. The methanol extract was evaporated to dryness, dissolved in methanol, and subjected to thin layer chromatography (TLC) with silica gel 60 (Merck) (developed with chloroform: methanol=4:1). The yellow pigments were separated into 3 spots having Rf values of 0.93, 0.62 and 0.30 by the TLC. The yellow (to orange) pigment at the Rf value of 0.30 which was the strongest spot was scraped off from the TLC plate, extracted with a small amount of methanol, loaded on a Sephadex LH-20 column for chromatography [30 cm.times.3.0 cm (.phi.)] and developed and eluted with methanol to give 4 mg of pure products. The yellow (to orange) pigment obtained was scarcely soluble in organic solvents other than methanol and easily soluble in water, suggesting that it might be a carotenoid glycoside. This was also supported from the molecular weight of 892 by FD-MS spectrum (the mass of this pigment was larger than that of zeaxanthin (described hereinafter) by the mass of two glucose molecules). When this substance was hydrolyzed with 1N HCl at 100.degree. C. for 10 minutes, zeaxanthin was obtained. Accordingly, the pigment was analyzed by acetylation according to the usual method. A large excess of acetic anhydride was added to the pigment dissolved in 10 ml of pyridine and the mixture was stirred at room temperature and left standing overnight. After the completion of reaction, water was added to the mixture and chloroform extraction was carried out. The chloroform extract was concentrated and loaded on a silica gel column [30 cm .times.3.0 cm (.phi.)] and the pigment was eluted with chloroform. The pigment gave a .sup.1 H-NMR spectrum identical with the tetraacetyl derivative of zeaxanthin-.beta.-diglucoside [Helvetica Chimica Acta, 57, 1641-1651 (1974)], identifying itself as zeaxanthin-.beta.-diglucoside (its structure being illustrated below).

The yield of the pigment was 1.1 mg/g dry weight. It had a solubility of at least 2 mg in 100 ml of water and methanol, water being a better solvent than methanol. It had low solubilities in chloroform and acetone; it dissolved only 0.5 mg in 100 ml of these solvents. ##STR1##

Experimental Example 4: Analysis of the metabolic intermediates of carotenoids

A hybrid plasmid (designated as pCAR16) was constructed by inserting the 1-6009 fragment of SEQ 1D NO:13, which was obtained by deletion to nucleotide position 6010 from the HindIII site (the 3' end) of the 6918 bp fragment containing the yellow pigment-synthesizing genes (DNA sequence SEQ ID NO:13) using deletion kit for kilo-sequence. pCAR16 contains genes crtB, E, I, X, and Y. Various deletion plasmids were constructed, as shown in Table 1, based on pCAR16 and the aforementioned hybrid plasmid pCAR25 (containing all the six genes).

The number within parentheses behind the name of respective restriction enzymes represents the number of the first nucleotide of its recognition site. The nucleotide numbers correspond to those in FIG. 1 and SEQ ID:NO:13. Analyses of metabolic intermediates of the carotenoid synthesis were performed using E. coli JM109 transformed with these deletion plasmids [referred to hereinafter as E. coli (name of plasmid)].

The cells of E. coli (pCAR25de1B) (exhibiting orange) harvested from a 3 liter culture in 2.times.YT medium were extracted twice with 400 ml portions of acetone at low temperature, concentrated, then extracted with chloroform:methanol (9:1) and evaporated to dryness. The extract was dissolved and subjected to silica gel column chromatography [30 cm.times.3.0 cm (.phi.)]. After the column was washed with chloroform, an orange band was eluted with chloroform:methanol (100:1). This pigment was dissolved in ethanol, recrystallized at low temperature to give 8 mg of pure products. Analyses by its UV-visible absorption, .sup.1 H-NMR, .sup.13 C-NMR and FD-MS (m/e 568) spectra revealed that this substance had the same planar structure as zeaxanthin (.beta.,.beta.-carotene-3,3'-diol). It was then dissolved in diethyl ether: isopentane: ethanol (5:5:2), and the CD spectrum was measured. As a result, it was found that this substance had a 3R,3'R-stereochemistry [Phytochemistry, 27, 3605-3609 (1988)]. Therefore, it was identified as zeaxanthin (.beta.,.beta.-carotene-3R,3'R-diol), whose structure is illustrated below. The yield was 2.2 mg/g dry weight. This substance corresponded to the yellow pigment having an Rf value of 0.93 in Experimemal Example (1). ##STR2##

The cells of E. coli (pCAR16) (exhibiting orange) harvested from a 3 liter culture in LB medium were extracted 3 times with 500 ml portions of cold methanol at low temperature and the methanol extract was further extracted with 1.5 liter of hexane. The hexane layer was concentrated and subjected to silica gel column chromatography [30 cm.times.3.0 cm (.phi.)]. Development and elution were conducted with hexane:ethyl acetate (50:1) to collect an orange band. The orange fraction was concentrated and recrystallized from ethanol to give 8 mg (reduced weight without moisture) pure products. This substance was presumed to be .beta.-carotene from its UV-visible absorption spectrum, and the molecular weight of 536 by FD-MS spectrum also supported this presumption. Upon comparing this substance with an authentic sample (Sigma) of .beta.-carotene by .sup.13 C-NMR spectrum, all the chemical shifts of carbons were identical. Thus, this substance was identified as .beta.-carotene (all-trans-.beta.,.beta.-carotene, whose structure was illustrated below). It was also confirmed by a similar method that E. coli (pCAR16de1B) accumulated the same .beta.-carotene as described above. Its yield was 2.0 mg/g dry weight, which corresponded to 2-8 times (per dry weight) of the total carotenoid yield in carrot (Kintokininjin) cultured cells described in Soshikibaiyou (The Tissue Culture), 13, 379-382 (1987). ##STR3##

The cells of E. coli (pCAR16de1C) (exhibiting red) harvested from a 3 liter culture in LB medium were extracted once with 500 ml of cold methanol at low temperature, and the precipitate obtained by centrifugation was extracted again with 1.5 liter of chloroform. The chloroform layer was concentrated and subjected to silica gel chromatography [30 cm.times.3.0 cm (.phi.)]. Development and elution were conducted with hexane:chloroform (1:1) to collect a red band. This fraction was concentrated. This substance was presumed to be lycopene from its UV-visible absorption spectrum, and the molecular weight of 536 by FD-MS spectrum also supported this presumption. Upon comparing this substance with the authentic sample (Sigma) of lycopene by .sup.1 H-NMR spectrum, all of chemical shifts of hydrogens were identical. When, this substance and the authentic sample were subjected to TLC with silica gel 60 (Merck) [developed with hexane:chloroform (50:1)] and with RP-18 [developed with methanol:chloroform (4:1)], the displacement distances of these samples were completely equal to each other. Thus this substance was identified as lycopene (all-trans-.psi.,.psi.-carotene, whose structure was illustrated below). It was also confirmed by a similar method that E. coli (pCAR-ADE) and E. coli (pCAR-ADEF) accumulated the same lycopene as described above. The yield of the former was 2.0 mg/g dry weight, which corresponded to 2 times (per dry weight) of the total carotenoid yield in a hyperproduction mutant of carrot (Kintokininjin) culture cells described in Soshikibaiyou (The Tissue Culture), 13, 379-382 (1987). ##STR4##

The cells of E. coli (pCAR-AE) were harvested from a 1.5 liter culture in 2.times.YT medium were extracted twice with 200 ml portions of acetone, concentrated, extracted twice with 100 ml portions of hexane, and evaporated to dryness. This was subjected to silica gel chromatography [30 cm.times.3.0 cm (.phi.)]. Development and elution were conducted with hexane:chloroform (1:1) to collect a band which had a strong UV absorption, and it was confirmed to be phytoene by its UV absorption spectrum. It was further subjected to Scphadex LH-20 column (Pharmacia) chromatography [30 cm.times.3.0 cm (.phi.)]. Development and elution were conducted with chloroform:methanol (1:1) to give 4 mg of pure products. The comparison of the .sup.1 H-NMR spectrum of this substance with the .sup.1 H-NMR spectra of trans- and cis-phytoene (J. Magnetic Resonance, 10, 43-50 (1973)) showed this substance to be a mixture of the trans- and cis-isomers. Isomerization from trans-isomer to cis-isomer hardly occurs, and thus it was judged that such a mixture was produced as a result of cis-trans isomerization in the course of the purification. It was therefore concluded that the original phytoene was the cis-type phytoene (15,15'-cis-phytoene, whose structure was shown below). It was also confirmed by a similar method that E. coli (pCAR25de1D) accumulated the same phytoene as described above. ##STR5##

Experimental Example 5: Identification of carotenoid biosynthesis genes

From the facts that E. coli (pCAR25) produced zeaxanthin-diglucoside and that E. coli (pCAR25de1B) harboring a plasmid in which crtX had been removed from pCAR25 accumulated zeaxanthin, it was found that the crtX gene encoded the glycosylation enzyme which was capable of convening zeaxanthin into zeaxanthin-diglucoside. Similarly, from the fact that E. coli (pCAR16de1B) harboring a plasmid in which crtZ had been removed from pCAR25de1B accumulated .beta.-carotene, it was found that the crtZ gene encoded the hydroxylation enzyme which was capable of convening .beta.-carotene into zeaxanthin. Similarly, from the fact that the E. coil (pCAR-ADE) harboring a plasmid in which crtY had been removed from pCAR16de1B accumulated lycopene, it was found that the crtY gene encoded the cyclization enzyme which was capable of convening lycopene into .beta.-carotene. E. coli (pCAR-ADEF) carrying both of the crtE, crtI and crtB genes required for producing lycopene and the crtZ gene encoding the hydroxylation enzyme synthesized only lycopene. This demonstrates directly that the hydroxylation reaction in carotenoid biosynthesis occurs after the cyclization reaction. Further, from the facts that E. coli (pCAR-ADE) accumulated lycopene and that E. coli (pCAR-AE) harboring a plasmid in which the crtI gene had been removed from pCAR-ADE accumulated phytoene, it was found that the crtI gene encoded the desaturation enzyme which was capable of convening phytoene into lycopene. E. coli (pCAR-A) carrying only the crtE gene and E. coli (pCAR-E) carrying only the crtB gene were not able to produce phytoene. These results indicate that both the crtE and crtB genes were required for producing phytoene in E. coli.

A 3 ml overnight culture of E. coli JM101 (pCAR-A) or E. coli JM101 (pCAR-E) was inoculated to 50 ml LB medium and was further grown at 37.degree. C. for 3 hours. The cells were harvested and resuspended in 1 ml of 0.1 M Tris-HCl (pH 8.0) and broken by a French press at a pressure of 500 bar. The suspension was treated with DNase (20 .mu.g/ml) for 15 minutes on ice, centrifuged at 10,000 g for 3 minutes, and the supematant was assayed for geranylgeranyl pyrophosphate (GGPP) synthase activity. The assay mixture contained in 0.5 ml: 5 .mu.mol ATP, 3 .mu.mol MnCl.sub.2, 2 .mu.mol MgCl.sub.2, 20,000 dpm of trans,trans-2-.sup.14 C-famesyl pyrophosphate (FPP, specific activity 1.88 GBq/mmol; purchased from Amersham), and 400 .mu.l of the supematant. The mixture was incubated at 30.degree. C. for two hours and then HCl was added to a final concentration of 2 M to hydrolyze the prenyl pyrophosphates in the mixture into their alcohols. After keeping the mixture at room temperature for 30 minutes, the prenyl alcohols were extracted by hexane containing 10% diethylether and separated by HPLC with a Spherisorb ODS-1 5.mu.g, 25 cm column (Phase Separations Ltd.) using 10% (v/v) water in methanol as the solvent at a flow rate of 1 ml/min. Radioactivity in the eluent was determined on-line with a radioactivity monitor. In vitro GGPP synthase activity was calculated as the percentage of radioactivity in GGPP vs. the sum of radioactivity in FPP and GGPP. As shown in Table 2, E. coli strains carrying the crtE gone (pCAR25 and pCAR-E) had an elevated activity to convert FPP into GGPP, which indicates that crtE encodes GGPP synthase.

The activity of the reaction next to GGPP synthesis in carotenoid biosynthesis, the conversion of GGPP into phytoene, was assayed for the E. coli strains in Table 2 in a similar manner, but for some unknown reason the activity was not detected for any of the strains. Therefore, two plasmids for transforming Agrobacterium tumefaciens containing only either crtB or crtE gene were constructed by inserting either the Eco52I (4926) - EcoRI (6009) fragment or the KpnI (1) - BstEII (1235) fragment from the DNA sequence of SEQ ID NO:13, respectively, into the filled-in StnaI-SstI large fragment of a binary vector pBI121 (Clontech). The resultant plasmids, pBICRT-B and pB[CRT-E, had the .beta.-glucuronidase gene replaced by respective crt genes. A. tumefaciens LBA4404 transformants carrying either of these plasmids were grown overnight at 28.degree. C. in a 10 ml medium containing yeast extract 5 g/l, tryptone 20 g/l, MgSO.sub.4 .multidot.7H.sub.2 O 10 g/l, adjusted to pH 7.5. The culture was inoculated into 50 ml of the same fresh medium and grown for further 4 hours, and the cells were harvested and resuspended in 1 ml of 0.1 M Tris-HCl pH 8.0. The cells were broken by a French press at 500 bar, treated with DNase (20 .mu.g/ml) for 15 minutes on ice, and assayed for phytoene synthase activity. The assay mixture contained in 0.5 ml: 5 .mu.mol ATP, 3 .mu.mol MnCl.sub.2, 2 .mu.mol MgCl.sub.2, 500,000 dpm of R-2-.sup.14 C-mevalonic acid (specific activity 56.7 mCi/nmol; purchased from Amersham), 200 .mu.l of a Fttsarium SG4 extract, and 200 .mu.l of the cell homogenate. The Fusarium extract, which converts the .sup.14 C-mevalonic acid into .sup.14 C-GGPP, was prepared as described by Sandmann and Albrecht [Z. Naturforsch., 45c, 487-491, (1990)]. After incubating at 30.degree. C. for two hours, 2.5 ml of methanol containing 6% (v/v) KOH was added and the lipids in the assay mixture were saponified at 65.degree. C. for 15 minutes in the dark, and non-saponifiable lipids were extracted with hexane containing 10% diethylether and analyzed by HPLC as in the GGPP synthase assay described above except that the solvent used was acetonitrile:methanol:2-propanol (85:10:5, v/v).

In the assay mixture containing the homogenate ofA. turnefaciens (pBICRT-B), radioactive 15-cis as well as all-trans phytoene were detected, which indicates crtB encodes phytoene synthase that can convert GGPP directly into phytoene. The homogenate of A. tumefaciens (pBICRT-E) was not able to form phytoene. From these analyses described above, all of the six crt genes have been identified and the biosyhthetic pathway of carotenoids have also been clear. These results are listed in FIG. 2.

E. coli (pCAR25de1E) accumulated no detectable carotenoid intermediate, while E. coli (.pCAR25de1A) and E. coli (pCAR-CDE) were able to produce a small amount of carotenoids. That is to say, E. coli (pCAR25de1A) and E. coli (pCAR-CDE) produced 4% of zeaxanthin-diglucoside and 2% of .beta.-carotene as compared with the E. coli (pCAR25) and the E. coli (pCAR16de1B), respectively. These results are consistent with the observation that E. coli without any crt genes was capable of synthesizing GGPP at a low level (see Table 2). It is not clear at this moment, however, why phytoene was not detected in E. coli (pCAR-E) carrying the crtB gene.

As described above, in the present invention the detailed biosynthetic pathway of carotenoids including general and famous carotenoids such as lycopene, .beta.-carotene and zeaxanthin and a water soluble carotenoid zeaxanthin-diglucoside were for the first time elucidated, and a gene cluster useful for their biosynthesis was obtained for the first time. In this connection, lycopene, .beta.-carotene and zeaxanthin which were produced by the genes in the aforementioned Experimental Examples were stereochemically identical with those derived from higher plants [T.W. Goodwin: "Plant Pigments", Academic Press (1988)].

Zeaxanthin-diglucoside has been isolated from a plant [Pure & Appl. Chem., 47, 121-128 (1.976)] but its isolation from microorganisms has not been reported before.

Experimental Example 6: Synthesis of carotenoids in Zymomonas

Zymomonas mobilis is a facultative anaerobic ethanol-producing bacterium. It has a higher ethanol producing rate than yeast (Saccharomyces cerevisiae), so that it is preferable as a fuel alcohol-producing bacterium in future. Zymomonas has a special metabolic pathway, Entner-Doudoroff pathway, instead of the glycolytic pathway, and do not produce carotenoids. In order to add further values to this bacterium, the carotenoid biosynthesis genes were introduced into Zymomonas.

The 7.6 kb fragment containing the DNA sequence of SEQ ID NO:13 was cut out from the hybrid plasmid pCAR1 by KpnI digestion and treated with DNA polymerase I (Klenow enzyme). The fragment thus treated was ligated to the EcoRV site of the cloning vector pZA22 for Zymomonas [see Agric. Biol. Chem., 50, 3201-3203 (1986) and Japanese Patent Laid-Open Publication No. 228278/87] to construct a hybrid plasmid pZACAR1. Similarly, the 1-6009 fragment of the DNA sequence SEQ ID N0:13 was cut out from pCAR16 by KpnI/EcoRI digestion and treated with DNA polymerase I (Klenow enzyme). The fragment thus treated was ligated to the EcoRV site of pZA2,2 to construct a hybrid plasmid pZACAR16. The orientation (as in SEQ ID NO:13) of the inserted fragments in these plasmids was opposite to the orientation of the Tc.sup.r gene. These plasmids were introduced into Z. mobilis NRRL B-14023 by conjugal transfer with the helper plasmid pRK2013 (ATCC 37159) and stably maintained in this strain. Z. mobilis NRRL B-14023 in which pZACA. R1 or pZACAR16 had been introduced formed a yellow colony and produced zeaxanthin-diglucoside in an amount of 0.28 mg/g dry weight or .beta.-carotene in an amount of 0.14 mg/g dry weight, respectively. Carotenoids were thus successfully produced in Zymomonas by the carotenoid biosynthesis genes of the present invention.

Experimental Example 7: Synthesis of .beta.-carotene and lycopene in Saccharomyces cereviceae

Saccharomyces is a yeast widely utilized for fermentation and bakery. No natural species of this yeast is reported to synthesize carotenoids. In the following experimental example, the DNA sequences 1, 3, 4 and 5 of the present invention were introduced and expressed in Saccharomyces cereviceae to produce .beta.-carotene and lycopene in the transformants.

The genomic DNA of Saccharomyces cereviceae RC7 (.alpha.leu2 ura3 trp1) was partially digested with Sau3AI into about 10 kb fragments and inserted into the BamHI site of pBR322 to construct a yeast genomic DNA library. A clone containing the GAL7 gene was obtained from this library by screening with a 21met synthetic oligonucleotide probe complementary to the nucleotide 1-21 of the coding sequence of the GAL7 gene [Molec. Cell. Biol., 6, 246-256 (1986)]. The genomic DNA fragment in this clone was partially digested with Sau3AI into about 600 bp fragments and subcloned into PUC19. A clone containing the GAL7 promoter fragment was obtained using the same 21mer probe described above. The 5' upstream of the promoter fragment was deleted up to about 270 bp upstream of the transcription start by exonuclease III (Takara Shuzo, Co.) and a BamHI linker was added there. Similarly, the 3' downstream of the promoter fragment was deleted up to 6 bp upstream of the transcription start and a BgIII linker was added. The 300 bp BamHI-BgIII fragment thus obtained, which contained the GAL7 promoter, was inserted between the BamHI and BgIII sites of a yeast expression vector YEp13K (European patent application 339 532; referred to as EP 339 532 below). The resultant plasmid was linearized with BgIII, filled in with the Klenow fragment and self-ligated after the addition of a HindIII linker to construct the plasmid pGAL7pro.

A clone containing the GAL7 terminator was obtained from the PUC19 subclone library described above by screening with a 30mer synthetic probe complementary to the nucleotide sequence 30 bp upstream of the stop codon of GAL7. A 420 bp HindIII-SalI fragment containing the terminator was excised from this clone and inserted between the HindIll and SalI sites of pUC19 to construct the plasmid pGAL7term.

Another yeast genomic library was constructed in the same manner as described above from the genomic DNA of Saccharomyces cereviceae S288C ATCC 26108. A clone containing the PGK gene was obtained from this library by screening with a 28mer synthetic oligonucleotide probe complementary to the nucleotide 1-28 of the coding sequence of the PGK gene [Nucleic Acids Res., 10, 7791 (1982)]. A 2.9 kb fragment containing the PGK gene was obtained from the plasmid in this clone by HindIII digestion and was inserted into pUC19. The 2.3 kb HindIII-SalI fragment containing the PGK promoter and the first half of the coding sequence was isolated from this plasmid and inserted between the HindIII and SalI sites of pBR322 to construct pBRPGK. A series of deletions were constructed by an exonuclease III digestion of the SalI-linearized pBRPGK. The digested DNA was filled in with the Klenow fragment, ligated with HindIII linkers, digested with HindIII, and cloned into the HindIII site of pUC19. A plasmid carrying a PGK fragment deleted up to 7 bp upstream of the transcription start was selected. This plasmid was partially digested with HindIII to cut only the site upstream of the PGK promoter, filled in with the Klenow fragment and self-ligated after the addition of a BglII linker to construct the plasmid pPGKpro.

Plasmid pPG2 (EP 339 532), which carries the PGK terminator, was cut with SmaI, ligated with HindIII linkers and digested with HindIII and SalI to isolate the 0.38 kb HindIII-SalI fragment containing the PGK terminator. This fragment and the 1.6 kb BglII-HindIII fragment containing the PGK promoter excised from pPGKpro were ligated between the BglII and SalI sites of YEp13K to obtain an expression vector Y-PGK-PT.

A clone containing the HIP1 gene was obtained from the S288C genomic library described above by screening with a 30met synthetic oligonucleotide probe complementary to the nucleotide 46-75 of the coding sequence of the HIP1 gene [Gene, 38, 205-214 (1985)]. A 1 kb fragment containing the HIP1 terminator was isolated from the plasmid in this clone with BamHI and SalI and was inserted between the BamHI and SalI sites of pBR322 to construct the plasmid pHIPT.

pCAR16 was digested with Eco52I, filled in with the Klenow fragment, digested with EcoRI and the 1083 bp (Eco521)-EcoRI fragment containing the crtB gene was isolated. (Eco52I) represents a filled in Eco52I site. In the following document, parenthesized restriction sites represent the filled in sites. pUC19 was cut with SphI, filled in and digested with EcoRI. The (Eco52I)-EcoRI fragment containing the crtB gene was ligated between the (SphI) and EcoRI sites of pUC].9 thus obtained to construct a plasmid pcrtB. The 901 bp TaqI-EcoRI fragment containing the most part of the crtB gene was excised from pcrtB and ligated between the SphI and EcoRI sites of M13mp18 (Takara Shuzo Co.) together with the synthetic oligonucleotide fragment shown below to construct mp18-crtB. ##STR6## mp18-crtB was cut with EcoRI, filled in with the Klenow fragment and self-ligated after the addition of a XhoI linker to obtain mp18-crtBXhoI. A BamHI-HindlII fragment containing the GAL7 promoter was excised from the plasmid pGAL7pro described in above (1), and the HindlII-XhoI fragment containing the crIB gene was excised from mp18-crtBXhoI. These fragment were together ligated between the BgllI and XhoI sites of YEp13K to construct Y-GAL7-crtB, on which the GAL7 promoter is linked to the 5' of the crIB gene via the HindlII site.

The 1.87 kb SnaBI-EcoRV fragment containing the crtI gene was excised from pCAR16 and inserted between the (SphI) and SmaI sites of pUC18 to construct pcrtI. The orientation of the crtI gene on pcrtI was such that the EcoRI and HindlII sites of pUC18 were upstream and downstream, respectively, of the gene. pcrtI was linearized with HindlII and about 200 bp sequence was deleted with exonuclease III. After ligating HindlII linkers, the 1.7 kb EcoRI-HindlII fragment containing the crtI gene was isolated and inserted between the EcoRI and HindlH sites of pUC18 to construct pcrtld. A XbaI site was created at 18 bp upstream of the transcription start of crtI on pcrtld by an in vitro mutagenesis [Bio/Technology, 2, 636-639 (1984)] using a synthetic oligonucleotide shown below to obtain pcrtlXbaI.

5' ACGACTCATCTAGAAGGAGCGACTAC 3 ' (SEQ ID NO: 16 )

The 1.0 kb XhoI-(HindlII) fragment containing the GPD promoter was excised from a yeast expression vector pAX50F2 (lip 339 532), and the (XbaI)-HindIII containing the crtl gene was excised from pcrtIXbaI. These fragments were together ligated between the XhoI and HindIII sites of Bluescript II SK(+) (Stratagene) to construct pGPD-crtI. A unique SalI site is conveniently located upstream of the GPD promoter on pGPD-crtI. A SaII-HindIII fragment containing the GPD promoter-crtI gene was excised accordingly, and a HindIII-SalI fragment containing the GAL7 terminator was excised from the plasmid pGAL7term described in above (1). These fragments were together inserted in the SalI site of pUC19 to obtain pGPD-crtIT, on which the GPD promoter is linked to the 5' of the ertl gene via the XbaI site and the GAL7 terminator to the 3' via the Hitcliff site.

The 1.1 kb Spel-PvuI fragment containing the crtE gene was excised from pCAR16 and made blunt-ended with mung bean nutlease, and inserted into the filled in Xbal site of pUC19 to construct portE. The orientation of the crtE gene on pcrtI was such that the EcoRI and Salt sites of pUC19 were upstream and downstream, respectively, of the gene. pcrtE was digested with SacI and BamHI, treated with mung bean nuclcase to delete the sequence up to 9 bp upstream of the crtE coding region, and self-ligated. The 0.94 kb EcoRI-SalI fragment containing the crtE gene was excised from this plasmid, made blunt-ended with mung bean nuclease, and inserted into the HindIII site of pUC19 after the addition of a HindIII linker to obtain pcrtE-H. The 0.94 kb HindIII fragment containing the crtE gene was excised from pcrtE-H and inserted into the HindIII site of the expression vector Y-PGK-PT described above to construct Y-crtE, on which the PGK promoter is linked to the 5' and the PGK terminator to the Y of the crtE gene via the HindlII site.

The 1.3 kb EcoRV-BamHi fragment containing the crtY gene was excised from pCAR16 and inserted between the SmaI and BamHI sites of pUC19 to construct pettY. The 1190 bp Nrul-BamHI fragment containing the most part of the crtY gene was excised from pcrtY and ligated into the BamHI site of pUC19 together with the synthetic oligoDNA fragment shown below to construct pcrtY-BS. ##STR7## pcrtY-BS has a new SmaI site 13 bp upstream of the transcription start of the crtY gene. The 1.2 kb SmaI-BamHI fragment containing the crtY gene was excised from pcrtY-BS, and the 1.5 kb BamHI-(HindIII) fragment containing the ADH1 promoter was excised from a plasmid pYADH (EP 339 532) carrying the promoter. These fragments were together ligated into the BamHI site of pUC19 to obtain pADH-crtY. The orientation of the crtY gene on pADH-crtY was such that the SalI site of pUC19 was upstream of the ADH1 promoter. The 2.7 kb SalI-BamHI fragment containing the ADH1 promoter-crtY gene was excised from pADH-crtY and the BamHI-SalI fragment containing the HIP1 terminator was excised from the plasmid pHIPT described in above (1). These fragments were ligated into the SalI site of pUC19 to obtain 19-crtY, on which the ADH1 promoter is linked to the 5' of the crt[gene and the HIP1 terminator is linked to the Y via the BamHI site.

A vector plasmid Y-crtBEIY for conferring the ability of carotenoid (such as .beta.-carotene and lycopene) synthesis to yeast was constructed as follows.

The BamHI-Xhol fragment containing the GAL7 promoter-criB gene was excised from Y-GAL7-crtB and inserted between the BgllI and XhoI sites of Y-crtE to obtain Y-crtBE. Then, the SalI fragment containing the GPD-promoter-crtI gene-GAL7 terminator was excised from pGPD-crtlT and inserted into the SaII site of Y-crtEB to obtain Y-crtEBI. Finally, the SalI fragment containing the ADH1 promoter-crtY gene-HIP1 terminator was excised from 19-crtY and ligated with Y-crtEBI partially digested with SalI to obtain Y-crtBEIY as shown in FIG. 3. The plasmid has the LEU2 gene as a selection marker.

In Y-crtBEIY, the criB, E, I and Y genes are regulated by the GAL7, PGK, GPD and ADH1 promoters, respectively. Of these four promoters, the GAL7 promoter is a galaclose inducible promoter and the others are constitutive ones. Accordingly, in order to express all the four promoters at the same time, it is necessary to add galactose to the culturing medium. The yeast strain to be transformed should therefore be able to grow well on a medium containing galactose. One example of such yeast strain is Saccharomyces cereviceae RC7 (ct leu2 ura3 trpl), which was used as the host in the following experiment.

Y-crtBEIY was introduced to the yeast strain by the lithium method [J. Bacteriol., 153, 163 (1983)] and transformants were selected on a medium without leucine. One transformant was arbitrarily chosen and designated as BCAR1. BCAR1 was cultured for 3 days at 30 .degree. C. in 200 ml of a selective medium SG-Leu containing galactose [0.67% amino acid-free yeast nitrogen base (Difco), 2% galactose, 20 mg/l arginine-HCl, 20 mg/l histidine-HCl, 30 mg/l isoleucine, 30 mg/1 lysine-HCl, 20 mg/l methionine, 50 mg/l phenylalanine, 200 mg/l threonine, 30 mg/l tyrosine, 20 mg/l tryptophan, 20 mg/l adenine sulphate, 20 mg/l uracil, pH 5.8]. Carotenoids were extracted from the yea-st according to Johnson, E.A. et aL [J. Gen. Microbiol., 115, 173-183 (1979)]. The harvested cells were broken with glass beads and the carotenoids were extracted by shaking with a twenty volume of acetone. The acetone extract (60 ml) was filtered and the carotenoids were extracted twice with 25 ml of hexane. After extracting water- soluble impurities by shaking with 25 ml of water for five times, the hexane layer was dehydrated, concentrated and evaporated to dryness. The carotenoids were dissolved in a small volume of chloroform and analyzed with HPLC using a Nova-pak HR C.sub.18 6.mu. column (3.9.times.300 mm; Waters). The carotenoids were eluted with acetonitrile:methanol:2-propanol (90:6:4 ) at a flow rate of 1 ml/min. Synthetic all-trans-.beta.,.beta.-carotene (Sigma) and lycopene from tomato (purchased from Sigma) were used as standards. Upon the HPLC analysis of the yeast extract, a peak was eluted at the same retention time as the standard .beta.-carotene. The peak was recovered and its absorbance spectrum was analyzed to confirm it to be .beta.-carotene. Lycopene was detected and identified in the yeast extract in a similar way. The yield of .beta.-carotene was 91 .mu.g per g dry weight of the cells.

In another experiment, BCAR1 was further transformed with a plasmid vector pTRP56 (Clontech), which has the TRP1 gene, by the lithium method, and transformants were selected on a medium without leucine and tryptophane to obtain a yeast transformant BCAR2. BCAR2 was examined for its .beta.-carotene production in the same way as for BCAR1 except that the medium was devoid of tryptophane. BCAR2 produced .beta.-carotene at a yield about 1.2 times higher than BCAR1.

Microorganism relating to the present invention is deposited at Fermentation Research Institute, Japan as follows: